Tired office worker works as 'AI bot nanny'
According to a report published in mid-June by the Work AI Research Institute (WAI) of Glean company (USA) with the University of Notre Dame, Stanford University, and University of California Berkeley, office workers spend an average of 6.4 hours per week providing context, checking output, correcting errors, and handling bot errors.
The report uses the term "nanny bot" to describe this underappreciated but essential task. The research team surveyed 6,000 full-time workers in the US, UK, and Australia, mainly working with computers or digital tools from December 2025 to January 2026. The results showed that 87% of participants used AI and 75% said this technology helps work more efficiently, but only 13% rated their organization as significantly better thanks to AI.
Rebecca Hinds, head of WAI, said that most of the "lost" productivity comes from work that employees never thought they would have to do. On the Cognitive Revolution podcast last week, Hinds described the job of bot nanny as boring, tiring, unrewarded or unappreciated, unrecognized, and unmeasured.
According to Business Insider, this pressure seems to be negatively affecting employee morale. WAI's report shows that people who have to spend too much time taking care of bots are 73% more likely to find another job. "People who have to shoulder the burden without being recognized or rewarded will feel exhausted and dissatisfied. After that, they begin to refine their resumes," the research team commented.
Victor Dibia, an AI researcher at Microsoft, said last year that trying to keep up with new breakthroughs causes "AI fatigue," to the point where he forgets his laptop password. On
According to digital workplace researcher Elizabeth Marsh, workers are suffering from "emotional exhaustion" due to work overload, and AI could exacerbate that condition. She explained to the Telegraph that having to monitor and check tools consumes a lot of energy and causes mental fatigue. For example, when AI offers many options, employees must verify again, especially with "seemingly correct" results. Over time, a series of small decisions accumulates into great pressure.
Dissatisfaction and fatigue also stem from employees having to become "middlemen" for technologies that do not work well together. They spend time transferring information between disparate AI systems, correcting errors and providing context where they should already be available.
"I have one tool to help me consider technical decisions, another tool to help me draft and summarize. I keep switching between them, double-checking every little detail. Instead of working faster, my mind starts to get confused. Not physically tired, just... overloaded," a senior engineering manager said in a study published in the Harvard Business Review in March.
This study shows that AI can cause employees to fall into brain fry. The conclusion was made after the authors, including experts from consulting firm Boston Consulting Group and the University of California, Riverside, surveyed 1,500 workers. About 14% of participants said they had felt mentally exhausted from having to deal with AI tools. They describe many harmful effects such as feeling "buzzed" or foggy, headaches, difficulty concentrating, slower decision making, and more errors.
In some cases, employees are asked to automate parts of their jobs that they enjoy. Hinds gives the example of a customer service employee who enjoys building relationships but is gradually forced to become an AI assistant supervisor. Losing what brings joy and meaning to work is "very dangerous".
According to research teams from WAI and universities, companies cannot simply deploy more AI tools. The organizations that benefit the most often invest in work around AI like helping employees access the right context, teaching them how to use technology effectively, and setting clear standards for how to apply AI well.
AI is considered to allow employees to do more work in less time. But according to Julie Bedard, CEO of Boston Consulting Group, if it pushes workers to burnout, organizations need to reconsider this perspective. "We need to redesign the way we work, not simply keep what we did yesterday and just bring in AI," she told Harvard Business Review.
Bedard and his colleagues' research also shows how leadership and employee training play an important role. Brain overload is less likely to occur in employees whose managers proactively use AI appropriately.
Experts warn that if no changes are made, businesses will continue to pay the price with high AI nanny costs and the departure of employees who are tired of cleaning up after chatbots.
Thu Thao compiled